Most micro data are collected by national statistical offices and, in turn, require confidential treatment so that published data tables are usually restricted to aggregate figures (e.g., household income or industrial production). Such data are, of course, also available in disaggregated form: For example, income may be differentiated by region or occupation, and production may be available for a single industry or for a single region. Additionally, tables of two or more dimensions are often provided, which, for example, may display income by region and occupation or production by industry and region. Certain cells in these tables, however, cannot be revealed due to the intrinsic possibility of identifying individuals or, more frequently, specific firms. A famous example is automobile production within a certain region; all production is typically due to one large enterprise, and the corresponding publication of the production of local automobile manufactures will clearly compromise this large firm's confidentiality.
展开▼
